RAG Evals with RAGAS: Build, Score, and Fix Your Pipeline

What is this?

A 2.5-hour hands-on workshop(in-person and online) where you build a RAG pipeline from scratch, evaluate it with RAGAS — the industry-standard eval framework — and learn to measure whether your pipeline is actually
working.

---
What you will do:

- Build a small RAG pipeline

• Grade RAG answers by hand first — so you understand what "good" means
• Run RAGAS to automate that grading at scale
• Deliberately break your pipeline and watch the scores drop
• Fix it and watch the scores recover

---
What you will leave with:

• A working eval pipeline you can drop into any project
• Understanding of the key metrics that matter: faithfulness, answer relevancy, context recall
• A repeatable workflow: change something → run evals → know if it got better
• A GitHub repo to show on your portfolio

---
Who this is for:

• You have heard of RAG but never properly evaluated one
• You built a RAG system and you are going on vibes about whether it works
• You want something concrete and practical, not a lecture

---
Prerequisites:
Basic Python. Laptop with Python 3.12, Docker Desktop, and an Anthropic API key (free tier works). No prior eval experience needed.

Online: Will share a google meet link.
Organizers: https://www.linkedin.com/in/kumar-satyam-4b871951/